I work for a very small company that builds and customizes a reporting web application as part of a package that another large firm sells. It's currently written in Perl, so it would be a Perl product that's sold again and again.
What's special about our product is that it's not a product, per se. It's more of a service that happens to be written in Perl. In other words, we're a consulting firm in disguise. And, frankly, that's how you're going to have to think of yourself. While you may be able to ride the gravy train of a certain product for a few years, you're not going to build your business unless you're a domain expert first and a producer second.
There's a lot of really bad programmers out there*, but it's not like they have their abilities tattooed on their foreheads. You need to have all the credentials in the world in order to be a recognized domain expert. For example, Stonehenge is a well-known Perl consulting firm. Why? Because Randal Schwartz and brian_d_foy are there! If you want to hire someone, you want to go with the best you can afford. Stonehenge has two of the TOP names in the Perl world. In other words, you have to compete with that if you want to grow your business.
Now, it's not all doom-and-gloom. You can go away from the Perl-centric idea and just build a good web-based product to market. For example, that's what Basecamp does. I bet you can't figure out what language they're using. I'll give you a hint - it's not Perl.
*: According to some estimates, the top 1% of the world's programmers are more efficient than the other 99% combined. Figuring out which side of that line you fall on is tough. But, the nice thing is that it's easy to go from one side to the other, if you're willing to devote yourself to it.
My criteria for good software:
- Does it work?
- Can someone else come in, make a change, and be reasonably certain no bugs were introduced?
What does it take for an application to succeed in the marketplace that you're describing? Well, it needs to be installable on Windows in the same fashion that Firefox is installable on Windows. You download something, double-click on something, click "Yes" a few times, double-click on something else and you're up and running.
This takes a lot of infrastructure on the Windows machine. While it's all possible to do with Perl, it makes a lot more sense to do it in a language that already has all that infrastructure in place, like C# or Java. Why compete in a space that's already saturated?
Instead, it's much better to work within a space that doesn't have that cost of entry, like rich web applications. Have you used Basecamp? It's a damn slick application - I like it ... a lot.
As for "only way out" ... I make a fine living doing exactly what you're describing, and so does nearly every other serious Perl/Python/Ruby developer. Just because it's not flashy doesn't mean it's not a good thing to do.
My criteria for good software:
- Does it work?
- Can someone else come in, make a change, and be reasonably certain no bugs were introduced?
I wouldn't be at all surpised if that statement were true, but I also wouldn't be surprised if the top 5%-1% had a combined efficincy greater than the top 1%. The bottom end of the scale has negative efficiency.
If you're calling yourself a "Perl programmer", you're already limiting your career path. Chad Fowler tells you all about this in his recent Perlcast interview where he says that you should make yourself useful in a certain problem area regardless of the tools. Be a programmer who knows Perl along with everything else you know, but also know other things. Don't get tied up just being a programmer either.
Have you looked at Bricolage, RT, or Movable Type? Those are all Perl products. Don't fool yourself into thinking that retail scales better than consulting. More revenue doesn't mean more profit or more return for your investment.
brian d foy
Subscribe to The Perl Review
You have a couple of options:
- Retail: needs a flashy GUI more than anything; not Perl's strongest point. Other than that, for retail it doesn't matter what language you write it in. Just realise that it's very, very hard to get on the shelves. And even harder to get a decent margin. Expect big up-front costs, and lots of risk, and a big advertisement budget. Republishing is an alternative strategy in this market, but expect to be sucked dry. If you make it in, and have something popular, you can make a decent living by counting on the volume of the market. You'll likely need to go global.
- Small business: also a big market in terms of volume. Won't pay unless it's dead simple to install and maintain, or it's customised to their specific needs, or it's a commodity (like Windows or Office).
- Enterprise: lots of cash, lots of requirements. I don't think this is a market where you can sell anything off-the-shelf.
In my experience, the consumer market isn't worth the pain. Support alone could kill you. I haven't directly done anything with big enterprises, but I suspect it's not feasible to get a foot in the door without expensive consultants and an impressive portfolio of buzzwords.
In the end, I think the small business segment is the most exciting: there are plenty opportunities for small shops selling intranets, web sites and so on. You'll be customizing your base product for each client, which technically makes you a consultant, but there's no reason why you shouldn't base this on a common framework that you develop for your particular niche. It gives you at least three income streams:
- you sell your application framework
- you bill your customizations by the hour
- you sell a support contract
I don't think there's anything wrong with developing web sites. You can deploy those skills with equal ease into selling intranets, and I think those are just software like any other, but with a lot of management benefits over desktop applications. If you view the web browser as your GUI toolkit, perl comes out very, very strong.
I personally think Perl is best suited to quick and dirty custom programming. For instance, I wrote a script yesterday to merge the results from a four-part poll, remove extreme results, and tabulate votes per person and votes per item. The script took maybe an hour to write, but writing the same script in any other language would have taken significantly longer, and processing the data by hand would have taken ages. Perl to the rescue!
The module CPAN::DistroBuilder does exactly what you said about packing the modules version that you need. PAR also could help in this issue, as well.
---------------------------------
"You have enemies? Good. That means you've stood up for something, sometime in your life." - Sir Winston Churchill
Don Wilde
"There's more than one level to any answer."
I always feel employed/consulting is not a scalable business proposition.
You're right to a large extent, but this touches on something that I always wonder about: why is the (western) world so obsessed with growth and scaling? Everyone seems to be looking for "fire & forget" businesses that they can set up then sit back while the product yields money. It's probably just the way I am wired, but I would rather earn my money by providing a service and get the benefits of becoming more efficient at that service, rather than selling "product".
Sorry, I'm probably off-topic here, and I'm certainly not answering your question, but I needed to get that off my chest.
I'll leave you with a thought to mull over: by the time you're fifty, like I'll be in three months, just about everything (technical) that you've learned, except the wisdom you've gained the hard way, will be completely obsolete, and change will be happening at such a pace that keeping your skill sets applicable will be daunting. If you _haven't_ got something that provides a sustaining chunk of money by that time, you're going to be in big trouble. In rare cases, like Y2K or COBOL consulting, you can make tasty gravy with no-longer-current skills, but those are rare cases.
OTOH, business skills are a sustainable skillset with continuing market value. Whether you use them for your own profit, or use them in service to a larger organization, they're a lot more transportable than programming or engineering skills. In my own company, the bills are paid by products we've already built as we work on next year's. In my day job, it's mostly my team leadership that makes a difference. In both cases, skill in leveraging other people is the key to success, far more than the elegant web apps I personally create.
I personally would like to retire in a few more years, long before I'm 65. 401K's aside, having a business with enough sustained revenue and growth to sell is really the only way to do that. I may be too old by that time to really push the Ferrari I'm going to buy, but I sure as heck am going to enjoy it. :D
Don Wilde
"There's more than one level to any answer."
I like to think of this as I'm in this s**t for life. Build a service, build a product based out of a service, fine tune my services to sell my product better. And so on! (Horrible example, please excuse)
Even Microsoft I´m sure they are concerned about how to sell Windows as a service, not as a product (they are developing a MSOffice version to be sold as service already, OpenOffice/Google is going this way too). Oracle database is gone, Pg/MySQL is out there, thats why they are buying companies that sell more services than software, like ERP ones, Peopleware, and developing service based applications. Think of Google: does they seel anything at all?
Source-code seeing or decompile and pirates are problem for everybody. And even if they were not, by the time you release a new version of something you have been developing for years, there are plenty os competitors in the same niche.
Companies like Basecamp, Google, Yahoo and many other Web 2.0 enabled simply don´t care about the language - and so don´t the user. In all the companies, the software is only the way to guve the final user the real revenue - information faster and better, collaboration, etc.
Diego de Lima
Well, good thing using Perl then :) No waste of time doing all hard work. Just do you're idea, and not control every damn byte like you have to do with C/C++...
That's true for code that does "general computing", like creating hashes or a brand new technique to query a database. But I think we're not going to see so soon companies putting strategic and high value business process rules into a product using programming languages like Perl or Java, which are (relative) easy to get the source code
.---------------------------------
"You have enemies? Good. That means you've stood up for something, sometime in your life." - Sir Winston Churchill